Case study
Prose Captain
Prose Captain: an AI tool for creating, editing and optimising content that writes from a specific company's knowledge, using RAG and a Pinecone vector database.

For Prose Captain, we designed and built an AI content creation tool. At its heart is an editor where generating articles and posts goes hand in hand with editing them, while AI assistants work on clients' knowledge bases and optimise texts for search engines.
- Industry
- Startups
- Product type
- AI content creation tool
- Market
- Startups / content creation
- Headless CMS
- Product Design
- Sztuczna Inteligencja - AI
- Content editor combining AI generation and editing
- Articles, blog posts and other content formats
- AI assistants working on clients' knowledge bases
- RAG (Retrieval-Augmented Generation) and a Pinecone vector database
- Automatic SEO optimisation of generated content
An editor where AI writes from company knowledge and a human refines the text.

Goals
01
One place to work on content
Combine generating, editing and optimising content in a single editor.
02
Content from company knowledge
Make the AI write from the knowledge of a specific industry and company, not generic data.
03
Search visibility
Create search-friendly content without a separate optimisation step.
Process


What we delivered
A content creation product: an AI editor, a client knowledge layer and SEO optimisation.
- Content editor - generating articles, blog posts and other content formats in one interface, with editing and adjusting of the results.
- Assistants on client data - AI assistants working on client databases, tailored to the knowledge and needs of different industries and companies.
- RAG and Pinecone - vector database search that gives language models context from the client's materials.
- SEO optimisation - automatic preparation of search-friendly content.
- Product design - UX/UI design of the whole tool.
Challenge: AI without context writes in generalities
A language model without context writes in generalities. To be accurate, company content has to draw on the knowledge of a specific industry and company.
Prose Captain had to combine generation with editorial work: AI text is a starting point that needs to be easy to correct and adapt. On top of that, the content had to be search-friendly from the start.
AI was meant to speed up writing without sacrificing content quality.


Solution: a content editor with AI and company knowledge
The core of the product is an editor that combines AI capabilities with an interface for everyday work on text.
Articles, blog posts and other content formats are created in the editor. Generated text can be edited and adjusted straight away, in the same place.
Context from your own data
AI assistants work on client databases. With RAG and a Pinecone vector database, the language model receives pieces of a specific company's knowledge, so content fits its industry and needs. The system also optimises texts for SEO automatically.


Results: generation, editing and SEO in one place
Companies got a tool where generating, editing and optimising content happen in one place, based on their own knowledge.
Combining generation with easy editing is meant to shorten the time needed to prepare long-form content while keeping its quality, and SEO optimisation helps it reach audiences through search engines.


Written by Radosław Kołacki Published Updated
A lot of content to produce? AI can write from the company's knowledge.
We'll show you how to connect your company's materials with a content editor and SEO optimisation.
Blog
Learn from our experts
Practical articles on researching, designing and delivering digital products.
Mobile apps
Custom mobile app: stages from diagnosis to development
A custom mobile app project has four phases: diagnosis, design, development and optimisation. We describe the decision made in each and what remains afterwards.
12 Jan 2025 · Radosław Kołacki
Read more
-
Mobile apps
How much does it cost to build a mobile app?
17 Jul 2024 · Radosław Kołacki
-
Business
Event Storming: what it is and how to run a workshop
15 Jul 2024 · Radosław Kołacki
-
Business and UX metrics
Engagement rate in a SaaS product: how to calculate it, with 2 examples
13 Dec 2023 · Radosław Kołacki